Version 2 (#1)

* Added an Utf8Alphabet implementation

* Completely revamped implementation. Dropped hierarchical logic to reduce storage requirements

* Added tests for the Levenshtein implementation

* Fixed Utf8Alphabet

* Added support for Damerau-Levenshtein

* Fixed algorithm

* Removed Damerau-Levenshtein attempt

* CS

Author: Toflar

README.md

@@ -21,14 +21,16 @@ composer require toflar/state-set-index
 ```php
 namespace App;
 
-use Toflar\StateSetIndex\Alphabet\InMemoryAlphabet;
+use Toflar\StateSetIndex\Alphabet\Utf8Alphabet
+use Toflar\StateSetIndex\DataStore\InMemoryDataStore;
 use Toflar\StateSetIndex\StateSet\InMemoryStateSet;
 use Toflar\StateSetIndex\StateSetIndex;
 
 $stateSetIndex = new StateSetIndex(
     new Config(6, 4),
-    new InMemoryAlphabet(),
-    new InMemoryStateSet()
+    new Utf8Alphabet(),
+    new InMemoryStateSet(),
+    new InMemoryDataStore()
 );
 
 $stateSetIndex->index(['Mueller', 'Müller', 'Muentner', 'Muster', 'Mustermann']);
@@ -44,15 +46,20 @@ you want to index and or search.
 ## Customization
 
 This library ships with the algorithm readily prepared for you to use. The main customization areas will be
-the alphabet (both the way it maps characters to labels) as well as the state set storage, if you want to make the index
+the alphabet (both the way it maps characters to labels) and the state set storage, if you want to make the index
 persistent. Hence, there are two interfaces that allow you to implement your own logic:
 
 * The `AlphabetInterface` is very straight-forward. It only consists of a `map(string $char, int $alphabetSize)` method 
   which the library needs to map characters to an internal label. Whether you load/store the alphabet in some 
-  database is up to you. The library ships with an `InMemoryAlphabet` for reference and simple use cases.
-* The `StateSetInterface` is more complex but is essentially responsible to load and store information about the 
-  state set of your index. Again, whether you load/store the state set in some
-  database is up to you. The library ships with an `InMemoryStateSet` for reference and simple use cases.
+  database is up to you. The library ships with an `InMemoryAlphabet` for reference and simple use cases. You don't 
+  even need to store the alphabet as we already have one with the UTF-8 codepoints, that's what `Utf8Alphabet` is 
+  for. In case you don't want to customize the labels, use `Utf8Alphabet`.
+* The `StateSetInterface` is responsible to load and store information about the state set of your index. Again, 
+  how you load/store the state set in some database is up to you. The library ships with an `InMemoryStateSet` 
+  for reference and simple use cases and tests.
+* The `DataStoreInterface` is responsible for storing the string you index alongside its assigned state. Sometimes 
+  you want to completely customize storage in which case you can use the `NullDataStore` and only use the 
+  assignments you get as a return value from calling `$stateSetIndex->index()`.
 
 You can not only ask for the final matching results using `$stateSetIndex->findMatchingStates('Mustre', 2)` which is 
 already filtered using a multibyte implementation of the Levenshtein algorithm, but you can also access intermediary 

ecs.php

@@ -2,8 +2,9 @@
 
 declare(strict_types=1);
 
-use PhpCsFixer\Fixer\Import\NoUnusedImportsFixer;
-use PhpCsFixer\Fixer\Phpdoc\NoSuperfluousPhpdocTagsFixer;
+use PhpCsFixer\Fixer\ClassNotation\OrderedClassElementsFixer;
+use PhpCsFixer\Fixer\FunctionNotation\NativeFunctionInvocationFixer;
+use PhpCsFixer\Fixer\Operator\NotOperatorWithSuccessorSpaceFixer;
 use Symplify\EasyCodingStandard\Config\ECSConfig;
 use Symplify\EasyCodingStandard\ValueObject\Set\SetList;
 
@@ -23,5 +24,7 @@
     ]);
 
     // Always move private elements to the bottom
-    $ecsConfig->rule(\PhpCsFixer\Fixer\ClassNotation\OrderedClassElementsFixer::class);
+    $ecsConfig->ruleWithConfiguration(OrderedClassElementsFixer::class, ['sort_algorithm' => 'alpha']);
+    $ecsConfig->rule(NativeFunctionInvocationFixer::class);
+    $ecsConfig->skip([NotOperatorWithSuccessorSpaceFixer::class]);
 };

src/Alphabet/InMemoryAlphabet.php

@@ -14,31 +14,31 @@ public function __construct(
     ) {
     }
 
-    public function all(): array
+    public function add(string $char, int $label): self
     {
-        return $this->alphabet;
+        $this->alphabet[$char] = $label;
+
+        return $this;
     }
 
-    public function has(string $char): bool
+    public function all(): array
     {
-        return isset($this->alphabet[$char]);
+        return $this->alphabet;
     }
 
     public function count(): int
     {
-        return count($this->alphabet);
+        return \count($this->alphabet);
     }
 
     public function get(string $char): ?int
     {
         return $this->alphabet[$char] ?? null;
     }
 
-    public function add(string $char, int $label): self
+    public function has(string $char): bool
     {
-        $this->alphabet[$char] = $label;
-
-        return $this;
+        return isset($this->alphabet[$char]);
     }
 
     public function map(string $char, int $alphabetSize): int

src/Alphabet/Utf8Alphabet.php

@@ -0,0 +1,21 @@
+<?php
+
+namespace Toflar\StateSetIndex\Alphabet;
+
+class Utf8Alphabet implements AlphabetInterface
+{
+    /**
+     * @var array<int, array<string, int>>
+     */
+    private array $cache = [];
+
+    public function map(string $char, int $alphabetSize): int
+    {
+        if (!isset($this->cache[$alphabetSize][$char])) {
+            // +1 in order to never assign 0
+            $this->cache[$alphabetSize][$char] = (mb_ord($char, 'UTF-8') % $alphabetSize) + 1;
+        }
+
+        return $this->cache[$alphabetSize][$char];
+    }
+}

src/Config.php

@@ -10,13 +10,13 @@ public function __construct(
     ) {
     }
 
-    public function getIndexLength(): int
+    public function getAlphabetSize(): int
     {
-        return $this->indexLength;
+        return $this->alphabetSize;
     }
 
-    public function getAlphabetSize(): int
+    public function getIndexLength(): int
     {
-        return $this->alphabetSize;
+        return $this->indexLength;
     }
 }

src/DataStore/DataStoreInterface.php

@@ -0,0 +1,16 @@
+<?php
+
+namespace Toflar\StateSetIndex\DataStore;
+
+interface DataStoreInterface
+{
+    public function add(int $state, string $string): void;
+
+    /**
+     * Returns the matching strings per state. Key is the state and the value is an array of matching strings
+     * for that state.
+     *
+     * @return array<int,array<string>>
+     */
+    public function getForStates(array $states = []): array;
+}

src/DataStore/InMemoryDataStore.php

@@ -0,0 +1,30 @@
+<?php
+
+namespace Toflar\StateSetIndex\DataStore;
+
+class InMemoryDataStore implements DataStoreInterface
+{
+    /**
+     * @var array<int, array<string>>
+     */
+    private array $data = [];
+
+    public function add(int $state, string $string): void
+    {
+        $this->data[$state][] = $string;
+    }
+
+    public function all(): array
+    {
+        return $this->data;
+    }
+
+    public function getForStates(array $states = []): array
+    {
+        if ([] === $states) {
+            return $this->data;
+        }
+
+        return array_intersect_key($this->data, array_flip($states));
+    }
+}

src/DataStore/NullDataStore.php

@@ -0,0 +1,16 @@
+<?php
+
+namespace Toflar\StateSetIndex\DataStore;
+
+class NullDataStore implements DataStoreInterface
+{
+    public function add(int $state, string $string): void
+    {
+        // noop
+    }
+
+    public function getForStates(array $states = []): array
+    {
+        return [];
+    }
+}

src/Levenshtein.php

@@ -4,7 +4,7 @@
 
 class Levenshtein
 {
-    public static function distance(string $string1, string $string2, int $insertionCost = 1, $replacementCost = 1, $deletionCost = 1)
+    public static function distance(string $string1, string $string2, int $insertionCost = 1, $replacementCost = 1, $deletionCost = 1): int
     {
         $string1 = mb_convert_encoding($string1, 'ASCII', 'utf8');
         $string2 = mb_convert_encoding($string2, 'ASCII', 'utf8');

src/StateSet/CostAnnotatedStateSet.php

@@ -11,28 +11,13 @@ class CostAnnotatedStateSet
      */
     private array $set = [];
 
-    /**
-     * Key: State
-     * Value: Cost
-     * @return array<int, int>
-     */
-    public function all(): array
-    {
-        return $this->set;
-    }
-
-    public function states(): array
-    {
-        return array_values(array_keys($this->set));
-    }
-
     /**
      * Adds a state with a cost to this set.
      * If this sets already contains the given state with a higher cost, replaces it.
      */
     public function add(int $state, int $cost): void
     {
-        if (! isset($this->set[$state])) {
+        if (!isset($this->set[$state])) {
             $this->set[$state] = $cost;
             return;
         }
@@ -43,6 +28,16 @@ public function add(int $state, int $cost): void
         }
     }
 
+    /**
+     * Key: State
+     * Value: Cost
+     * @return array<int, int>
+     */
+    public function all(): array
+    {
+        return $this->set;
+    }
+
     public function mergeWith(CostAnnotatedStateSet $stateSet): self
     {
         $clone = clone $this;
@@ -53,4 +48,11 @@ public function mergeWith(CostAnnotatedStateSet $stateSet): self
 
         return $clone;
     }
+
+    public function states(): array
+    {
+        $states = array_values(array_keys($this->set));
+        sort($states);
+        return $states;
+    }
 }

src/StateSet/InMemoryStateSet.php

@@ -5,77 +5,25 @@
 class InMemoryStateSet implements StateSetInterface
 {
     /**
-     * Key: State
-     * Value: array<parent,mappedChar>
-     *
-     * @var array<int, array<int,int>>
+     * @param array $states array<int, bool>
      */
-    private array $states = [];
-
-    /**
-     * @var array<int, array<int>>
-     */
-    private array $children = [];
-
-    /**
-     * Key: State
-     * Value: Mapped char
-     * @var array<int, int>
-     */
-    private array $mappedChars = [];
-
-    /**
-     * Key: State
-     * Value: Matching strings
-     * @var array<int, array<string>>
-     */
-    private array $acceptedStrings = [];
+    public function __construct(
+        private array $states = []
+    ) {
+    }
 
-    public function add(int $state, int $parentState, int $mappedChar): self
+    public function add(int $state): void
     {
-        $this->states[$state] = [$parentState, $mappedChar];
-        $this->mappedChars[$state] = $mappedChar;
-        $this->children[$parentState][$state] = true;
-
-        return $this;
+        $this->states[$state] = true;
     }
 
     public function all(): array
     {
         return $this->states;
     }
 
-    public function getChildrenOfState(int $state): array
-    {
-        if (! isset($this->children[$state])) {
-            return [];
-        }
-
-        return array_keys($this->children[$state]);
-    }
-
-    public function getCharForState(int $state): int
-    {
-        if (! isset($this->mappedChars[$state])) {
-            throw new \LogicException('No mapped char for state. Check your alphabet!');
-        }
-
-        return $this->mappedChars[$state];
-    }
-
-    public function acceptString(int $state, string $string): self
-    {
-        $this->acceptedStrings[$state][] = $string;
-
-        return $this;
-    }
-
-    public function getAcceptedStrings(array $matchingStates = []): array
+    public function has(int $state): bool
     {
-        if ([] === $matchingStates) {
-            return $this->acceptedStrings;
-        }
-
-        return array_intersect_key($this->acceptedStrings, array_flip($matchingStates));
+        return isset($this->states[$state]);
     }
 }

src/StateSet/StateSetInterface.php

@@ -4,22 +4,12 @@
 
 interface StateSetInterface
 {
-    public function add(int $state, int $parentState, int $mappedChar): self;
-
-    public function getChildrenOfState(int $state): array;
-
-    public function getCharForState(int $state): int;
+    public function add(int $state): void;
 
     /**
-     * Accept a string with a given state.
+     * @return array<int>
      */
-    public function acceptString(int $state, string $string): self;
+    public function all(): array;
 
-    /**
-     * Returns the matching strings per state. Key is the state and the value is an array of matching strings
-     * for that state. If no argument is passed, the entire accepted strings dataset is returned.
-     *
-     * @return array<int,array<string>>
-     */
-    public function getAcceptedStrings(array $matchingStates = []): array;
+    public function has(int $state): bool;
 }